96-week resistance analyses of the STaR study: rilpivirine/emtricitabine/tenofovir DF versus efavirenz/emtricitabine/tenofovir DF in antiretroviral-naive, HIV-1-infected subjects

Abstract

Background:

STaR (GS-US-264-0110) was a 96-week phase 3b study evaluating the safety and efficacy of two single-tablet regimens, rilpivirine/emtricitabine/tenofovir DF (RPV/FTC/TDF) and efavirenz/emtricitabine/tenofovir DF (EFV/FTC/TDF) in treatment-naive, HIV-1-infected subjects.

Methods:

Genotypic analyses (population sequencing) of HIV-1 protease (PR) and reverse transcriptase (RT) were performed at screening; subjects with pre-existing resistance to study drugs were excluded. The protocol-defined resistance analysis population had genotypic/phenotypic analyses at failure and baseline for PR and RT.

Results:

Through week 96, the resistance analysis population included 24/394 subjects (6.1%) receiving RPV/FTC/TDF and 9/392 subjects (2.3%) receiving EFV/FTC/TDF. In the RPV/FTC/TDF arm, HIV-1 isolates from 21/394 subjects (5.3%) developed non-nucleoside reverse transcriptase inhibitor (NNRTI) and/or nucleoside reverse transcriptase inhibitor (NRTI) resistance mutations and 20/21 isolates had both NNRTI and NRTI genotypic and/or phenotypic resistance. In the EFV/FTC/TDF arm, isolates from 4/392 subjects (1.0%) developed NNRTI and/or NRTI resistance mutations. Resistance development after week 48 was infrequent (1.0% RPV/FTC/TDF; 0.3% EFV/FTC/TDF). When stratified by baseline HIV-1 RNA ≤  or >100 000 copies/ml, 9/260 (3.5%) versus 12/134 (9.0%) RPV/FTC/TDF-treated subjects and 3/250 (1.2%) versus 1/142 (0.7%) EFV/FTC/TDF-treated subjects developed resistant isolates, respectively. Pre-existing NRTI- and NNRTI-associated resistance mutations (not related to study drugs) did not impact treatment response to either regimen.

Conclusions:

Resistance development to RPV/FTC/TDF consisted of NNRTI and NRTI mutations and was more frequent than resistance development to EFV/FTC/TDF through week 96. Emergent resistance after week 48 was infrequent in both arms. Within the RPV/FTC/TDF arm, resistance development was more frequent in subjects with baseline HIV-1 RNA >100 000 copies/ml compared to baseline HIV-1 RNA ≤ 100 000 copies/ml.

Keywords:

• Rilpivirine
• Efavirenz
• Resistance
• E138K
• M184V
• M184I

Introduction

Rilpivirine (RPV)/emtricitabine (FTC)/tenofovir disoproxil fumarate (TDF) (Complera^®/Eviplera^®) is a well-tolerated single-tablet regimen (STR) therapeutic option for the treatment of HIV-1 infection. In Europe, RPV/FTC/TDF is indicated for the treatment of HIV-1 infected adults without known resistance mutations to the components of RPV/FTC/TDF and with a viral load ≤ 100 000 HIV-1 RNA copies/ml.¹ In the USA, RPV/FTC/TDF is approved for use in HIV-1-infected treatment-naive adults with baseline viral load ≤ 100 000 copies/ml and in virologically-suppressed adults to replace their current regimen.² In the phase 3 studies of the components of RPV/FTC/TDF [ECHO (Study TMC278-TiDP6-C209) and THRIVE (Study TMC278-TiDP6-C215)], RPV demonstrated non-inferior efficacy and improved safety and tolerability compared with EFV in treatment-naive patients.^3–5 However, there was an increase in the number of virologic failures and virologic failures that developed resistance in subjects with high baseline viral load treated with RPV compared to those treated with EFV.⁶ In the SPIRIT study (GS-US-264-0106), switching to the RPV/FTC/TDF STR from a boosted protease inhibitor-based regimen in virologically-suppressed patients maintained HIV suppression with low rates of virologic failure and resistance development.⁷

The Single-Tablet Regimen (STaR) study (GS-US-264-0110) was a randomized, open-label, 96-week study designed to directly compare the safety and efficacy of the two STRs RPV/FTC/TDF and EFV/FTC/TDF in treatment-naive HIV-1-infected adults. STaR was the first trial to evaluate the use of the RPV/FTC/TDF single-tablet regimen in treatment-naïve subjects as well as the first randomized study where all subjects are taking a complete one pill once-daily regimen without additional placebo pills. The primary endpoint of non-inferiority for the proportion of subjects achieving HIV-1 RNA < 50 copies/ml at week 48 for RPV/FTC/TDF versus EFV/FTC/TDF was met, with response rates of 85.8% versus 81.6%, respectively (4.1% treatment difference; 95% CI: ( − 1.1–9.2%).⁸ At week 96, the response rates were 77.9% versus 72.4% for RPV/FTC/TDF versus EFV/FTC/TDF, respectively (5.5% treatment difference; 95% CI: − 0.6–11.5%).⁹

Resistance analyses of the STaR study at week 48 demonstrated overall lower rates of resistance development in both treatment arms compared to the phase 3 studies of the individual components of these regimens, ECHO and THRIVE.^8,10 The proportion of subjects that developed resistant isolates was similar between arms in subjects with baseline viral load ≤ 100 000 copies/ml, but was higher in RPV/FTC/TDF-treated compared to EFV/FTC/TDF-treated subjects with baseline viral load >100 000 copies/ml. Most isolates with emergent resistance in the RPV/FTC/TDF arm had both nucleoside reverse transcriptase inhibitor (NRTI)- and non-nucleoside reverse transcriptase inhibitor (NNRTI)-associated resistance. NRTI-associated resistance was infrequent in the EFV/FTC/TDF arm. These trends in NNRTI and NRTI resistance in the STaR study were consistent with the earlier phase 3 studies of RPV as an individual agent.^6,10,11

Here, the genotypic and phenotypic characterization of emergent resistance through week 96 of the STaR study is reported. Analyses of baseline resistance mutations were also conducted to determine the effect of pre-existing NRTI- and NNRTI-associated mutations on treatment response to both study regimens through week 96.

Materials and Methods

STaR study design

Details of the STaR study have been previously published.⁸ Briefly, STaR (GS-US-264-0110; clinicaltrials.gov identifier: NCT01309243) was a phase 3b, randomized, open-label, multi-center, international, 96-week study evaluating the safety and efficacy of the RPV/FTC/TDF STR compared to the EFV/FTC/TDF STR in treatment-naive HIV-1 infected subjects. Subjects were randomized 1:1 to RPV/FTC/TDF or EFV/FTC/TDF. Eligibility criteria included screening HIV-1 RNA ≥ 2500 copies/ml, no prior ARV therapy, genotypic sensitivity to EFV, FTC, and TDF, and lack of the RPV mutations K101E/P, E138A/G/K/Q/R, Y181C/I/V, and H221Y. Randomization was stratified by screening HIV-1 RNA ( ≤ 100 000 or >100 000 copies/ml).

Virologic failure definition and resistance analysis population

Subjects who experienced either virologic non-response or virologic rebound, as defined below, were considered to have virologic failure. Virologic non-response was assessed at week 8 and was defined as having HIV-1 RNA ≥ 50 copies/ml and < 1 log₁₀ reduction from baseline at the week 8 visit, which was confirmed at the subsequent visit. Virologic rebound was defined as having two consecutive visits with HIV-1 RNA ≥ 50 copies/ml after achieving HIV-1 RNA < 50 copies/ml, or as having two consecutive visits with >1 log₁₀ increase in HIV-1 RNA from nadir. Subjects with virologic failure and HIV-1 RNA ≥ 400 copies/ml who were on study drugs were included in the resistance analysis population (RAP). The sample from the confirmation visit (if available) was analyzed for resistance development. In addition, subjects who were on study drugs, had not been analyzed previously, and had HIV-1 RNA ≥ 400 copies/ml at week 48, week 96, or their last visit (at or after week 8), were also analyzed for resistance.

Virology assessments

Viral load (HIV-1 RNA copies/ml) was assessed using COBAS AMPLICOR Monitor Version 1.5 (Roche Diagnostics, Basel, Switzerland). Two-sided P values were determined using Fisher's exact test.

At screening, pre-existing resistance in the protease (PR) and reverse transcriptase (RT) portion of the pol gene was assessed by genotype (population sequencing) using the GeneSeq^TM assay (Monogram Biosciences, South San Francisco, CA, USA). These screening data were used for baseline resistance analyses. For post-baseline resistance analyses of subjects in the RAP, PR/RT genotyping (population sequencing) and phenotyping were performed using the PhenoSense GT^TM assay (Monogram Biosciences).¹² Retrospective baseline PR/RT phenotypes were determined for those subjects in the RAP with emergent resistance.

FDA snapshot analysis

In the snapshot analysis, subjects who had their last available HIV-1 RNA < 50 copies/ml in the study visit window (week 48 or week 96), while on randomized treatment, were classified as virologic successes. Subjects who had their last available HIV-1 RNA ≥ 50 copies/ml in the study visit window, while on randomized treatment, subjects who discontinued study drug before or in the study visit window due to lack of efficacy, and subjects who discontinued study drug before or in the study visit window due to reasons other than adverse event with last available HIV-1 RNA on treatment ≥ 50 copies/ml were classified as virologic failures. Subjects who had missing data in the study visit window but were on study drug, or subjects who did not have on-treatment HIV-1 RNA data in the study visit window due to study drug discontinuation for any other reason other than AE/death or lack of efficacy and had their last available HIV-1 RNA on treatment of < 50 copies/ml were classified as having no data in the study visit window.

Results

Genotypic and phenotypic resistance development through week 96

Of the 786 randomized and treated subjects, a total of 33 subjects met the criteria for inclusion in the RAP (33/786, 4.2%). Isolates from a total of 24/394 (6.1%) and 9/392 (2.3%) subjects were analyzed in the RPV/FTC/TDF and EFV/FTC/TDF arms, respectively, with post-baseline genotypic and phenotypic data for PR and RT available for all subjects (Table 1).

Table 1. Development of genotypic resistance at week 48 and week 96

	Number of subjects (%)
	RPV/FTC/TDF (n = 394)			EFV/FTC/TDF (n = 392)
Resistance development category	Baseline–week 48	Week 48–week 96	Baseline–week 96	Baseline–week 48	Week 48–week 96	Baseline–week 96
Resistance analysis population	20 (5.1%)	4 (1.0%)	24 (6.1%)	7 (1.8%)	2 (0.5%)	9 (2.3%)
Subjects with data	20	4	24	7	2	9
Developed resistance mutations to study drugs	17 (4.3%)	4 (1.0%)	21 (5.3%)	3 (0.8%)	1 (0.3%)	4 (1.0%)
Baseline viral load ≤ 100 000 copies/ml	5/260 (1.9%)	4/260 (1.5%)	9/260 (3.5%)	2/250 (0.8%)	1/250 (0.4%)	3/250 (1.2%)
Baseline viral load>100 000 copies/ml	12/134 (9.0%)	0	12/134 (9.0%)	1/142 (0.7%)	0	1/142 (0.7%)
No change from baseline (primary mutations^a)	3 (0.8%)	0	3 (0.8%)	4 (1.0%)	1 (0.3%)	5 (1.3%)
Any NNRTI-R^b	16 (4.1%)	4 (1.0%)	20 (5.1%)	3 (0.8%)	1 (0.3%)	4 (1.0%)
E138K/Q	6 (1.5%)	4 (1.0%)	10 (2.5%)	0	0	0
E138K	5 (1.3%)	3 (0.8%)	8 (2.0%)	0	0	0
E138K/Q	1 (0.3%)	1 (0.3%)	2 (0.5%)	0	0	0
Y181C/I	8 (2.0%)	0	8 (2.0%)	0	0	0
Y181C	5 (1.3%)	0	5 (1.3%)	0	0	0
Y181I	3 (0.8%)	0	3 (0.8%)	0	0	0
V90I^c,^d	6 (1.5%)	2 (0.5%)	8 (2.0%)	0	0	0
K101E	5 (1.3%)	0	5 (1.3%)	0	0	0
H221Y	3 (0.8%)	1 (0.3%)	4 (1.0%)	0	0	0
M230I/L	2 (0.5%)	0	2 (0.5%)	0	1 (0.3%)	1 (0.3%)
V179I	2 (0.5%)	0	2 (0.5%)	0	0	0
A98G	1 (0.3%)	0	1 (0.3%)	0	0	0
L100I	1 (0.3%)	0	1 (0.3%)	0	0	0
K103N	1 (0.3%)	0	1 (0.3%)	1 (0.3%)	0	1 (0.3%)
V106I	1 (0.3%)	0	1 (0.3%)	0	0	0
V108I	1 (0.3%)	0	1 (0.3%)	0	0	0
Y188H/L	1 (0.3%)	0	1 (0.3%)	1 (0.3%)	0	1 (0.3%)
G190E/Q	0	0	0	1 (0.3%)	0	1 (0.3%)
P225H	1 (0.3%)	0	1 (0.3%)	0	0	0
F227C	1 (0.3%)	0	1 (0.3%)	0	0	0
Any NRTI-R^e	16 (4.1%)	4 (1.0%)	20 (5.1%)	1 (0.3%)	1 (0.3%)	2 (0.5%)
M184V/I	15 (3.8%)	4 (1.0%)	19 (4.8%)	1 (0.3%)	1 (0.3%)	2 (0.5%)
M184I	8 (2.0%)	4 (1.0%)	12 (3.0%)	1 (0.3%)	1 (0.3%)	2 (0.5%)
M184V	5 (1.3%)	0	5 (1.3%)	0	0	0
M184V/I	2 (0.5%)	0	2 (0.5%)	0	0	0
K65R/N	3 (0.8%)	0	3 (0.8%)	0	0	0
K65R	2 (0.5%)	0	2 (0.5%)	0	0	0
K65N	1 (0.3%)	0	1 (0.3%)	0	0	0
K219E	3 (0.8%)	0	3 (0.8%)	1 (0.3%)	0	1 (0.3%)
A62V	1 (0.3%)	0	1 (0.3%)	0	0	0
D67N	0	0	0	1 (0.3%)	0	1 (0.3%)
T69T/del	1 (0.3%)	0	1 (0.3%)	0	0	0
K70E	1 (0.3%)	0	1 (0.3%)	0	0	0
L74V	1 (0.3%)	0	1 (0.3%)	0	0	0
Y115F	1 (0.3%)	0	1 (0.3%)	0	0	0

Note:

a No change from baseline at NNRTI-R, NRTI-R, or primary PI-R sites

b Non-nucleoside reverse transcriptase inhibitor resistance (NNRTI-R) mutations are V90I, A98G, L100I, K101E/H/P, K103N/S, V106A/M/I, V108I, E138A/G/K/R/Q, V179D/F/I/L/T, Y181C/I/V, Y188L/C/H, G190A/E/Q/S, H221Y, P225H, fig227C, and M230I/L in RT

c Three subjects also developed V189V/I

d K238N also developed in one subject with V90I

e Nucleoside/nucleotide reverse transcriptase inhibitor resistance (NRTI-R) mutations are M41L, E44D, A62V, K65R/N, D67N, tbl69D, tb69 insertions/deletions; K70E/R, L74V/I, V75I, fig77L, Y115F, fig116Y, V118I, Q151M, M184V/I, L210W, tbl215Y/F, and K219Q/N/E/R in RT.

In the RPV/FTC/TDF arm, of the 24 subjects with isolates analyzed for resistance development, 21 (21/394, 5.3%) had emergent resistance to a study drug (Table 1). Seventeen subjects had isolates with emergent resistance through week 48, and four subjects had isolates with emergent resistance between week 48 and week 96. Overall, more subjects with baseline HIV-1 RNA >100 000 copies/ml (12/134, 9.0%) developed emergent resistance compared with subjects that had baseline HIV-1 RNA ≤ 100 000 copies/ml (9/260, 3.5%) (P = 0.03). However, after week 48, no subjects with baseline HIV-1 RNA >100 000 copies/ml met the criteria for resistance analysis; all four subjects that developed resistant virus after week 48 had baseline HIV-1 RNA ≤ 100 000 copies/ml (Table 2). Through week 96, isolates from 20 subjects developed NNRTI resistance mutations, most commonly E138K/Q (n = 10), Y181C/I (n = 8), V90I (n = 8), and K101E (n = 5), and most had multiple NNRTI resistance mutations, frequently as mixtures with wild type. Isolates from 20 subjects developed NRTI resistance mutations, most commonly M184V/I (n = 19), K65R/N (n = 3), and K219E (n = 3). The remaining three subject isolates in the RPV/FTC/TDF arm lacked emergent genotypic and phenotypic resistance.

Table 2. Genotypic and phenotypic profiles of subjects in resistance analysis population with treatment-emergent NRTI or NNRTI resistance from post-week 48 through week 96

					Drug resistance mutations^a		Drug susceptibility (fold-change from WT)^b
Treatment arm	Baseline HIV-1 RNA (copies/ml)	HIV-1 RNA at analysis (copies/ml)	Time of virology analysis	Failure type	NNRTI^c	NRTI^d	RPV	EFV	FTC	TFV
RPV/FTC/TDF	10 500	1980	Week 96 Retest	VR^e	E138K	(V118I), M184I	2.3	1.7	>112	0.47
	12 400	70 600	Week 72	VR	E138K/Q	M184I, (K219R)	3.5	2.1	>96	0.54
	50 700	798	Week 84 Retest	VR	V90V/I, E138K, H221H/Y	M184I	4.3	1.8	>81	0.44
	59 100	2750	Week 96	VR	V90I, E138K	M184I	2.3	1.4	>96	0.62
EFV/FTC/TDF	97 700	648	Week 96 Retest	VR	(V179D), M230M/L	M184M/I	2.3	7.2	2.4	0.77

Note: NNRTI = non-nucleoside reverse transcriptase inhibitor; NRTI = nucleoside reverse transcriptase inhibitor; VR = virologic rebound; WT = wild type

a Mutations in parentheses were present at baseline. All HIV-1 isolates were subtype B

b Bolded cells represent a fold-change value greater than or equal to the biological or clinical cutoff for each drug. Cutoffs in the PhenoSense GT assay are RPV: 2.0; EFV: 3.0; FTC: 3.5; TFV: 1.4–4

c Non-nucleoside reverse transcriptase inhibitor resistance (NNRTI-R) mutations are V90I, A98G, L100I, K101E/H/P, K103N/S, V106A/M/I, V108I, E138A/G/K/R/Q, V179D/F/I/L/T, Y181C/I/V, Y188C/H/L, G190A/E/Q/S, H221Y, P225H, fig227C, and M230I/L in RT. A cumulative list of mutations that developed is shown

d Nucleoside/nucleotide reverse transcriptase inhibitor resistance (NRTI-R) mutations are M41L, E44D, A62V, K65N/R, D67N, tbl69D, tb69 Insertions; K70E/R, L74I/V, V75I, fig77L, Y115F, fig116Y, V118I, Q151M, M184I/V, L210W, tbl215F/Y, and K219E/N/Q/R in RT. A cumulative list of mutations that developed is shown

e Virologic rebound was not confirmed. Last visit on study drug.

Within the RPV/FTC/TDF arm, 20/21 isolates with any emergent resistance developed both NNRTI and NRTI genotypic and/or phenotypic resistance. Twenty isolates had primary NNRTI resistance mutations and reduced susceptibility to RPV (mean of 14-fold compared to wild type). Most of these isolates also showed reduced susceptibility to one or more other NNRTIs: 17 to etravirine, 12 to nevirapine, and 7 to EFV. Three isolates had reduced susceptibility to RPV without reduced susceptibility to any other NNRTIs. All 21 isolates with any emergent resistance had genotypic and/or phenotypic resistance to FTC with a mean of 91-fold reduced susceptibility to FTC compared to wild type. Among these isolates, 19 of 21 developed an M184V/I substitution and reduced susceptibility to FTC and lamivudine. One isolate had phenotypic resistance to FTC and lamivudine, but did not have any NRTI-associated resistance mutations by population sequencing. One additional isolate developed phenotypic resistance to FTC and lamivudine with K65R, a mixture of a one amino acid deletion and wild type at tb69 (T69T/del), and K219K/E; this isolate also showed reduced susceptibility to tenofovir. Two other subjects developed virus with K65R or N mixed with wild type without phenotypic resistance to tenofovir. All other isolates remained phenotypically susceptible to tenofovir.

In the EFV/FTC/TDF arm, of the nine subjects with isolates analyzed for resistance development, four (4/392, 1.0%) had emergent resistance to a study drug. Three subjects had isolates with emergent resistance through week 48, and one subject developed emergent resistance between week 48 and week 96 with baseline HIV-1 RNA ≤ 100 000 copies/ml (Table 2). Through week 96, all four subject isolates developed NNRTI resistance mutations (1 each of K103N, Y188L, G190E/Q, and M230L). Two of these also developed M184I. The remaining 5 subject isolates in the EFV/FTC/TDF arm lacked emergent genotypic and phenotypic resistance.

Within the EFV/FTC/TDF arm, 4 isolates had phenotypic resistance to EFV with a mean of 58-fold reduced susceptibility to EFV. Among these, all also showed reduced susceptibility to one or more of the other NNRTIs: 4 to nevirapine and 2 to RPV. All remained susceptible to etravirine. Two isolates developed the M184I substitution in RT and one had reduced susceptibility to FTC (>80-fold); both isolates with M184I had reduced susceptibility to lamivudine. All isolates remained susceptible to tenofovir.

Analysis of baseline mutations and treatment response

Isolates from all 786 randomized and treated subjects showed genotypic sensitivity to FTC, TDF, EFV, and RPV at their screening visit as required by the enrollment criteria. One subject in each treatment arm had E138A in their HIV, which was allowed at the beginning of screening but was subsequently added to the list of exclusion mutations in a protocol amendment. Both of these subjects were virologic successes at week 96. Other primary resistance mutations were observed among subject isolates at rates of 12.7% for NNRTI-, 8.3% for NRTI-, and 2.5% for protease inhibitor (PI)-associated mutations for the overall study population. Subjects with pre-existing NNRTI-, NRTI-, and PI-associated resistance mutations in their HIV that were not excluded at study entry had virologic responses that were similar to the responses observed for the overall study population at week 96 (Table 3).

Table 3. Virologic outcome by FDA snapshot analysis for select baseline resistance mutations at week 96

	RPV/FTC/TDF (n = 394)				EFV/FTC/TDF (n = 392)
		Snapshot outcome at week 96				Snapshot outcome at week 96
Resistance mutations at baseline, % (n)	Baseline frequency, % (n)	Virologic suppression	Virologic failure	No data in window	Baseline frequency, % (n)	Virologic suppression	Virologic failure	No data in window
Overall study results	100% (394)	77.9% (307)	9.4% (37)	12.7% (50)	100% (392)	72.4% (284)	5.9% (23)	21.7% (85)
NNRTI-R^a	11.7% (46)	80.4% (37)	10.9% (5)	8.7% (4)	13.8% (54)	81.5% (44)	3.7% (2)	14.8% (8)
V90I	2.5% (10)	90.0% (9)	0	10.0% (1)	3.1% (12)	83.3% (10)	0	16.7% (2)
A98G	0	0	0	0	0.3% (1)	100% (1)	0	0
V106I	1.0% (4)	75.0% (3)	0	25.0% (1)	2.6% (10)	70.0% (7)	0	30.0% (2)
V108I	0.8% (3)	100% (3)	0	0	0.8% (3)	100% (3)	0	0
E138A	0.3% (1)	100% (1)	0	0	0.3% (1)	100% (1)	0	0
V179D/I/T	7.9% (31)	77.4% (24)	16.1% (5)	6.5% (2)	7.4% (29)	82.8% (24)	6.9% (2)	10.3% (3)
V189I^b	1.5% (6)	100% (6)	0	0	1.0% (4)	100% (4)	0	0
NRTI-R^c	8.1% (32)	68.8% (22)	15.6% (5)	15.6% (5)	8.4% (33)	75.8% (25)	3.0% (1)	21.2% (7)
M41L	0.8% (3)	66.7% (2)	0	33.3% (1)	0.5% (2)	50.0% (1)	0	50.0% (1)
E44D	0.3% (1)	0	100% (1)	0	0.3% (1)	0	0	100% (1)
A62V	0	0	0	0	0.3% (1)	100% (1)	0	0
D67N	0.5% (2)	0	50.0% (1)	50.0% (1)	0.3% (1)	100% (1)	0	0
T69D	0.3% (1)	0	0	100% (1)	0	0	0	0
V75I	0	0	0	0	0.3% (1)	100% (1)	0	0
V118I	5.8% (23)	78.3% (18)	8.7% (2)	13.0% (3)	6.9% (27)	77.8% (21)	3.7% (1)	18.5% (5)
L210W	0.3% (1)	100% (1)	0	0	0.3% (1)	100% (1)	0	0
K219Q/R	0.8% (3)	0	33.3% (1)	66.7% (2)	0.5% (2)	100% (2)	0	0
Primary PI-R^d	2.3% (9)	88.9% (8)	11.1% (1)	0	2.8% (11)	45.5% (5)	27.3% (3)	27.3% (3)
L33F	0.8% (3)	100% (3)	0	0	0	0	0	0
M46I/L	0.8% (3)	100% (3)	0	0	1.0% (4)	50.0% (2)	0	50.0% (2)
Q58E	0.5% (2)	100% (2)	0	0	0.5% (2)	50.0% (1)	50.0% (1)	0
V82A	0.3% (1)	100% (1)	0	0	0.5% (2)	0	50.0% (1)	50.0% (1)
L90M	0.3% (1)	0	100% (1)	0	1.5% (6)	33.3% (2)	50.3% (3)	16.7% (1)

Note:

a NNRTI resistance mutations (NNRTI-R) are V90I, A98G, L100I, K101E/H/P, K103N/S, V106A/I/M, V108I, E138A/G/K/Q/R, V179D/F/I/L/T, Y181C/I/V, Y188C/H/L, G190A/E/Q/S, H221Y, P225H, fig227C, and M230I/L in RT

b V189I was not included in the list of NNRTI-R mutations studied, but is an additional substitution possibly associated with NNRTI resistance in other studies of RPV.⁹

c NRTI resistance mutations (NRTI-R) are M41L, E44D, A62V, K65N/R, D67N, tbl69D, tb69 insertions, K70E/R, L74I/V, V75I, fig77L, Y115F, fig116Y, V118I, Q151M, M184I/V, L210W, tbl215F/Y, and K219E/N/Q/R in RT

d Primary protease inhibitor resistance mutations (PI-R) are D30N, V32I, L33F, M46I/L, I47A/V, G48V, I50L/V, I54L/M, Q58E, tbl74P, L76V, V82A/F/L/S/T, I84V, N88S, and L90M in protease.

Discussion

Overall rates of resistance development through week 96 in the STaR study were low (5.3% RPV/FTC/TDF; 1.0% EFV/FTC/TDF) with infrequent emergent resistance after week 48 (1.0% RPV/FTC/TDF; 0.3% EFV/FTC/TDF). Resistance development was lower in both arms of the STaR study compared to the phase 3 studies conducted using the components of these two regimens (8.0% RPV+FTC/TDF; 3.1% EFV+FTC/TDF).² However, resistance development to at least one regimen component occurred more frequently with RPV/FTC/TDF than EFV/FTC/TDF through week 96.

Through week 96, the proportion of subjects with emergent resistance in their HIV was low in both arms among subjects with baseline HIV-1 RNA ≤ 100 000 copies/ml, while resistance development occurred more frequently among RPV/FTC/TDF-treated subjects with baseline viral load >100 000 copies/ml. Notably, no subjects with baseline viral load >100 000 copies/ml developed resistance after week 48 in either arm, demonstrating a reduced risk of virologic failure with resistance in this population once virologic suppression is achieved. There were no subjects that met the criteria for virologic nonresponse and resistance analysis after week 48 in either arm; all subjects that developed resistance after week 48 experienced virologic rebound after first achieving HIV-1 RNA < 50 copies/ml. These results are consistent with studies of RPV/FTC/TDF in virologically-suppressed subjects where emergent resistance was infrequent and independent of pre-treatment viral load.^7,13 Overall, these trends show that resistance development decreased with increasing duration of virologic suppression while taking RPV/FTC/TDF regardless of baseline viral load.

The emergent resistance patterns of RPV/FTC/TDF at week 96 are generally consistent with the week 48 results.^6,10,11 Most subjects in the RPV/FTC/TDF arm with virus that developed NNRTI resistance acquired multiple NNRTI mutations and complex patterns of resistance. No new RPV-associated mutations were observed after week 48 through week 96. The E138K/Q substitutions emerged most frequently through week 96, whereas Y181C/I were the most common emergent resistance mutations at week 48. Furthermore, through week 48, no subjects with baseline HIV-1 RNA ≤ 100 000 copies/ml developed resistance substitutions at position E138 in RT, while all four subjects that had their HIV analyzed for resistance after week 48 had baseline HIV-1 RNA ≤ 100 000 copies/ml and developed virus containing E138K or an E138K/Q mixture. Notably, the average RPV fold change for the four isolates with emergent resistance after week 48 was lower than for isolates with emergent resistance through week 48 (3.1-fold versus 17-fold, respectively). This likely reflects the more complex mutation patterns and presence of NNRTI mutations shown to confer higher levels of RPV resistance, such as Y181I, observed in isolates that developed resistance by week 48.^10,14 The reasons for these differences in mutation patterns may be related to the type of virologic failure or baseline viral load most frequently observed at each analysis time point (never suppressed and HIV-1 RNA >100 000 copies/ml through week 48 versus virologic rebound and HIV-1 RNA ≤ 100 000 copies/ml after week 48 through week 96).

Among those with any emergent resistance, more subjects in the RPV/FTC/TDF arm developed virus with both NRTI and NNRTI genotypic and/or phenotypic resistance. M184V/I substitutions were the most commonly-occurring NRTI mutations at both week 48 and week 96. All four of the isolates in the RPV/FTC/TDF arm that developed resistance after week 48 through week 96 had emergent M184I and phenotypic resistance to FTC. The combination of E138K+M184I that emerged in these isolates has been previously observed in subjects receiving RPV-containing regimens and has been shown to confer increased resistance to RPV compared to E138K alone in addition to FTC resistance.^{6,10,11,15–17} Other NRTI mutations including K65R or N, A62V, D67N, tbl69del, K70E, L74V, Y115F, and K219E occurred at very low frequencies of < 1% and were present mostly as mixtures with wild type. The prevalence of these mutations in HIV-1 infected individuals is generally low and may confer cross-resistance to multiple NRTIs when present in certain combinations.^18–21 With the exception of one isolate containing K65R, tbl69T/del, and K219K/E, the emergence of these mutations was not associated with reduced susceptibility to tenofovir or other NRTIs in the current study, thus the clinical relevance of these mutations is uncertain. Of note, several of these NRTI mutations were also observed at a low level in the phase 3 studies ECHO/THRIVE where they developed almost exclusively in subjects with baseline viral load >100 000 copies/ml.^6,11 In the STaR study, the emergence of the mutations in this group did not appear to be associated with baseline viral load as these mutations developed with similar frequencies in isolates from subjects with HIV-1 RNA > or ≤ 100 000 copies/ml.¹⁰

The relative frequencies of RPV resistance mutations observed in the STaR study were slightly different from the phase 3 studies ECHO/THRIVE.^6,10,11 In the phase 3 studies, E138K was the most frequent RPV-associated resistance mutation at both weeks 48 and 96 regardless of baseline viral load. Furthermore, Y181C occurred in only a small percentage of isolates at week 48 in those studies, mostly in subjects with baseline viral load >100 000 copies/ml, and was not observed after week 48. Y181I was not observed in any isolates tested.¹¹ Differences in dosing regimens and formulations between STaR and the phase 3 studies (one tablet once daily versus multiple pills, respectively) may have contributed to the variations in resistance development in these studies with the STR resulting in a moderately higher genetic barrier to resistance due to increased study drug compliance.

While RPV/FTC/TDF demonstrated non-inferior efficacy to EFV/FTC/TDF through week 96 of the STaR study, there was a greater incidence of resistance development as well as increased frequency of FTC- and RPV-associated resistance in the RPV/FTC/TDF arm compared to the EFV/FTC/TDF arm. Although the options for second-line regimens are more limited among subjects that experience virologic failure with resistance, there are numerous antiretroviral agents available that are efficacious against FTC- and RPV-associated resistance mutations including all drugs in the protease inhibitor and integrase strand transfer inhibitor classes, as well as TDF in most cases. Furthermore, the incidence of resistance development among subjects with baseline viral load ≤ 100 000 copies/ml on RPV/FTC/TDF in the STaR study was low (3.5%) and similar to resistance development in the overall study populations for other recommended antiretroviral regimens after 96 ♣weeks of treatment including elvitegravir/cobicistat/FTC/TDF (2.9%) and EFV/FTC/TDF (2.8%).²² This has led to the inclusion of RPV/FTC/TDF as a recommended initial antiretroviral regimen within the low viral load patient population in the US Department of Health and Human Services treatment guidelines.²³

The STRs in the STaR study demonstrated overall lower rates of resistance development in both treatment arms compared to the phase 3 studies of the individual components of these regimens through week 96. Resistance development was infrequent after week 48. The proportion of subjects that developed resistant isolates was low in both arms in subjects with baseline viral load ≤ 100 000 copies/ml, but was higher in RPV/FTC/TDF-treated compared to EFV/FTC/TDF-treated subjects with baseline viral load >100 000 copies/ml. While overall trends in NNRTI and NRTI resistance in the STaR study were consistent with the earlier phase 3 studies of RPV as an individual agent, slight variations were observed that may result from the use of the RPV/FTC/TDF STR.

Acknowledgements

The authors would like to thank all of the patients, study site staff, investigators, and Gilead staff who participated in the STaR study for their contributions to this work.

Disclaimer statements

Contributors

DPP, RK, and KLW analysed the data. TF was the medical monitor on the study and contributed to patient management and data analysis. MDM provided intellectual contributions to data analysis. DPP wrote the manuscript and RK, TF, MDM, and KLW reviewed and edited.

Funding

This work was sponsored by Gilead Sciences, Inc.

Conflicts of interest

All authors are employees and shareholders of Gilead Sciences, Inc.

Ethical approval

The clinical trial protocol was approved by the institutional review boards of the participating study sites. No additional ethical approvals were specifically warranted for the current analysis.